1. Field of the Invention
The present invention relates to an image sensing apparatus and, in particular, to an image sensing apparatus capable of high-speed image sensing suitable for scientific measurements such as measurement of ultra-high-speed moving objects like rockets, explosions, destructions, turbulent flows, discharge phenomena, chemical reaction phenomena, and motions of microorganisms under a microscope.
2. Description of the Prior Art
As this type of image sensing apparatus, there has conventionally been known, for example, an image converter type multi-framing camera (hereinafter, referred to as an image converter type camera). The image converter type camera adopts a system that, in an image sensing process in which an image converted into electrons at a photo-receptive area is irradiated onto a fluorescent surface with an electron gun, with the fluorescent surface divided into a plurality of areas, an electron flow is irradiated successively to the plurality of areas so that continuous image frames are displayed on one screen. This image converter type camera, in most cases, has a micro-channel plate type image intensifier (hereinafter, abbreviated as MCP type II) disposed behind the photo-receptive area or before the fluorescent surface, so that incident light is intensified. The frame rate of the image converter type camera is around 3.times.10.sup.7 frames/sec., and the number of image frames that can be continuously captured (number of continuous image frames) is 10 or so.
The present inventor has already provided a high-speed video camera comprising an MCP type II, wherein electric signals read in parallel from pixels are serially stored in memory provided outside the image sensor without forming an image, and the signals stored in the memory are formed into the image after the completion of image sensing, thus the video camera being enabled to implement a high-speed image sensing of 4,500 frames/sec. ("4500 frames/sec. high-speed video camera", Takeharu ETOH, the Journal of Television Society, Vol. 46, No. 5, PP. 543-545, 1992, in Japanese).
Further, the present inventor has proposed an image sensing apparatus, in Japanese Patent Laid-Open Publication No. HEI 5-336420, in which a plurality of electric signal storages are provided within each pixel, in which arrangement electric signals generated in a sensor in each pixel in response to brightnesses of incident light are accumulated in the electric signal storages within the pixels during an image sensing process, and read out after the completion of the image sensing process, thus the image sensing apparatus being enabled to implement high-speed image sensing.
Furthermore, as shown in FIGS. 32 and 33, there has been proposed an image sensing apparatus comprising an image sensor in which pixels 2 each having a signal storage CCD (charge coupled device) 1 are arranged in a matrix. The signal storage CCD 1 is folded by being bent four times and provided in a zigzag line within the pixel 2. During an image sensing process, one charge storage 3a out of the nine charge storages 3a to 3i for signal accumulation functions as a sensor that converts incident light into an electric signal, and the electric signal generated in the charge storage 3a is transferred and accumulated sequentially to the other charge storages 3b to 3i. After the completion of image sensing, the electric signals accumulated in the charge storages 3a to 3i of each pixel 2 are read out by signal read-out CCDs 4A, 4B ("THE STUDY OF A PHOTOSITE FOR SNAPSHOT VIDEO"; M. Elloum, E. Fauvet, E. Goujou, P. Gorria, G. Cathebras; 21st International congress on high speed photography and photonics (Aug. 29-Sep. 2, 1994), TECHNICAL PROGRAM & ABSTRACTS).
Further, U.S. Pat. No. 5,355,165, as shown in FIG. 34, describes an image sensing apparatus comprising an image sensor in which each pixel 6 comprises sensor 7 and signal storage CCD 9 having 5.times.4 charge storages 8, 8, . . . arranged in a matrix. Electric signals generated by the sensors 7 during an image sensing process are first transferred to f our charge storages 8 arrayed laterally (in the right-and-left direction) in the figure as indicated by arrow X. Next, with a bend of the direction in which the electric signals are transferred, they are transferred from the four charge storages 8 longitudinally (in the up-and-down direction) in the figure as indicated by arrow Y, thus being stored in the charges storages 8, respectively. During a read-out process, the electric signals stored in the charge storages 8 of the signal storage CCD 9 of each pixel 6 are transferred toward signal read-out CCD 10, longitudinally as indicated by the arrow Y, and further transferred by the signal read-out CCD 10 in the direction as indicated by the arrow X.
Generally, the image sensing apparatus for use of scientific measurement of motions is required to meet two conditions, that is, the capabilities of firstly implementing high-speed image sensing and secondly obtaining a sufficient number of consecutive images, or allowing use as a video camera.
As to the first condition, the present inventor has confirmed, as a result of questionnaires conducted on researchers who are users of image sensing apparatus for scientific measurement, that a frame rate of 10.sup.6 frames/sec., if obtained, satisfies the demands of most users (over 95%) as shown in FIG. 35 ("A survey by questionnaires on use of high-speed image sensing and a proposal of a 30,000,000 frames/sec. video camera"; Takeharu ETOH, Kohsei TAKEHARA, Midori KAWAJIRI; Proceedings of the 1993 comprehensive symposium on high-speed image sensing and photonics, pp. 109-112, 1993, in Japanese).
As to the second condition, since the reproduction rate of general video cameras is 25 to 30 frames/sec. but the minimum rate that can allow man to recognize a continuous motion is 4 to 5 frames/sec., reproducing 35 to 50 continuous frames at a reproduction rate of 4 to 5 frames/sec. allows a motion image of about 10 seconds to be obtained, although the motion in the image frames lacks in smoothness more or less, in which case the camera can be used as a video camera for scientific measurement use. Like this, image sensing apparatus for scientific measurement to be used as a high-speed video camera are required to be capable of obtaining a frame rate of 10.sup.6 frames/sec. and of obtaining a minimum number of continuous frames of around 35 to 50.
However, the aforementioned image converter type camera is indeed capable of obtaining a sufficient frame rate, but has a number of continuous frames as small as 10 or so, insufficient for use as a video camera. The image converter type camera also has disadvantages that the system results in large size, and that there occurs relative distortions between one and another of the continuous frames.
With 4,500 frames/sec. high-speed video cameras, as described above, the demand of users for the frame rate could not be satisfied sufficiently.
Further with the image sensing apparatus of Japanese Patent Laid-Open Publication No. HEI 5-336420, since a large number of transistors need to be controlled at one time in order to transfer electric signals to the electric signal storages during image sensing, large power consumption is involved in switching operations and besides noise generated in the switching affects the electric signals.
In an image sensing apparatus comprising the pixels 2 as shown in FIGS. 32 and 33, indeed there will occur neither lacks of intensity of light nor relative distortions between continuous frames, but the number of continuous frames is so low as 9 that the image sensing apparatus cannot be used as a video camera.
Also in this image sensing apparatus, since the charge storages 3a to 3i of the signal storage CCD 1 are arrayed in a zigzag line as described above, the drive circuit for operating the charge storages 3a to 3i of the signal storage CCD 1 becomes complex in construction, as another disadvantage. In more detail, as shown in FIG. 33, since a direction as indicated by arrow X1 in which electric signals are transferred from the charge storage 3a to the charge storage 3c and in which electric signals are transferred from the charge storage 3g to the charge storage 3i is opposite to another direction as indicated by arrow X2 in which electric signals are transferred from the charge storage 3d to the charge storage 3f, it is necessary to apply drive voltages having waveforms symmetrical to each other as shown by V1 and V2. Therefore, it is necessary that an electric wire 5a for applying a drive voltage to the charge storages 3a to 3c and the charge storages 3g to 3i, and an electric wire 5b for applying a drive voltage to the charge storages 3g to 3i are given by separate electric wires, and that the voltages of waveforms as shown by V1 and V2 are applied to the electric wires 5a and 5b, respectively. Besides, since a direction as indicated by arrow Y in which electric signals are transferred from the charge storage 3c to the charge storage 3d and from charge storage 3f to the charge storage 3g is perpendicular to the directions as indicated by X1 and X2, it is necessary to separately provide an electric wire 5c for applying a drive voltage to the charge storages 3d, 3f in addition to the electric wires 5a, 5b.
Further, since the signal storage CCD 1 is in a zigzag line as described before, there are bends in the direction in which electric signals are transferred. Electric signals will remain in the charge storages 3c, 3d, 3f and 3g placed at these bends in the transfer direction, which may cause deterioration of the quality of reproduced image.
Further in this image sensing apparatus, there is a need of providing signal read-out CCDs 4A within the photo-receptive area.
In general, the dimensions of the photo-receptive area of the pixel are about 10 mm.times.10 mm, whereas the number of pixels cannot be reduced below a certain level for the minimum resolution to be ensured. As a result, the area of one pixel cannot be increased over a certain level. Accordingly, in order to increase the number of signal storage CCDs and charge storages to be disposed within the pixel, it is necessary to simplify the construction within the photo-receptive area as much as possible so that a space for arranging the signal storage CCDs within each pixel is ensured. Therefore, as in the image sensing apparatus shown in FIGS. 32 and 33, if the signal read-outs CCD 4B are disposed within the photo-receptive area and if the drive circuit for the signal storage CCDs is complex, it would be difficult for the number of charge storages of the signal storage CCDs to be around 35 to 50 per pixel even if the dimensions and configuration of the CCDs are modified in various ways. Thus, the number of continuous frames necessary for use as the aforementioned high-speed video camera cannot be obtained.
Adopting the arrangement as shown in FIG. 34 allows the number of charge storages of the signal storage CCD 9 of each pixel 6 to be around 35 to 50 per pixel 6. However, since the electric signal generated in the sensor 7 is transformed first in a direction indicated by arrow X, and then in another direction indicated by arrow Y perpendicular to the X direction, the drive circuit for the signal storage CCD becomes complex as in the image sensing apparatus shown in FIGS. 32 and 33.
Also, since electric signals generated in the sensor 7 are transferred first in the direction as indicated by arrow X and then transferred in the direction as indicated by arrow Y, there are some abrupt bends in the direction in which the electric signal is transferred, as in the case of the arrangement shown in FIGS. 32 and 33. This poses a problem that electric charges will remain in the charge storages 8 placed at these bends, causing deterioration of the quality of reproduced image.
Meanwhile, in another example of this type of image sensing apparatus, as described in Japanese Patent Laid-Open Publication No. HEI 4-68876 filed by the present inventor, the image sensing apparatus comprises a beam split prism for splitting incident light into three, and a filter holder interposed between the emitting portion of the beam split prism and the image sensor, wherein a required filter can be fitted to the filter holders. This image sensing apparatus is enabled to perform color image sensing when a red-light reflecting filter, a green-light reflecting filter, and a blue-light reflecting filter are fitted to the filter holder. Without fitting the filters to the filter holder, if three image sensors are synchronized for image sensing, then a threefold resolution can be obtained as compared with the case of one camera. Further, if an image is captured with phase shifts of three cameras without fitting the filters, continuous frames in sets of three can be obtained and the speed, acceleration, and the like at each time point can be determined.
However, in the apparatus as described in Japanese Patent Laid-Open Publication No. HEI 4-68876, since incident light is split into three, or each one third the quantity of the incident light, fitting RGB filters in this arrangement would cause the intensity of light to be reduced to further one third, such that only one ninth the intensity of light would be reached on each image sensor.
The dichroic prism, on the other hand, selectively reflects red light, green light, and blue light, in which case the problem of lack in the intensity of light is improved. For this image sensing apparatus, however, it is impossible to implement high-resolution image sensing with three image sensors synchronized with one another, or to determine the speed, acceleration, and the like with phase shifts of the three cameras.